A forecast tells you what might sell. A plan has to decide what to make, buy, hold and promise when reality refuses to cooperate. Demand planning gets confused when teams treat forecasts as targets, inventory as demand, or accuracy measures as interchangeable. This page starts with the estimate and its error, then moves through lead time, service, order quantity and cross-functional decisions. It also separates the chance of avoiding a stockout from the share of units actually filled. Every formula needs a stated horizon, item set and denominator; examples below are illustrative, and cutoffs vary by organization.
🧭 Jump to a term
Estimate demand: Demand forecast · Demand planning · Constrained sales forecast · Seasonality · Demand sensing
Measure forecast performance: Forecast bias · WAPE · MAPE · FVA
Turn plans into stock decisions: Replenishment lead time · Cycle service level · Order fill rate · Economic order quantity (EOQ)
Prioritize and align: ABC analysis · XYZ analysis · S&OP
⭐ Know these first
Start with Demand forecast, Forecast bias, WAPE, Replenishment lead time, S&OP. Then follow the grouped learning order below.
📎 How to read this page
What it means gives the precise meaning. Operator translation gives the version you might hear in a real ecommerce meeting. In real life shows an illustrative example. Watch out and the confusion boxes show where a familiar term can mislead.
📈 Read the relationships first
These combinations are diagnostic hypotheses, not proof of causality. Compare the same period and scope, then investigate the mechanism.
Forecast bias ↑ + WAPE stable
Usually means: errors remain similar in size but lean more consistently high or low. Check next: signed error by product, horizon and override source.
Forecast demand ↑ + constrained sales flat
Usually means: supply or channel capacity may be limiting what can be fulfilled. Check next: on-hand stock, inbound timing, production capacity and lost-sales evidence.
Cycle service level ↑ + fill rate flat
Usually means: stockouts may be rarer while their size or duration remains material. Check next: units short per stockout and the service metric’s time boundary.
Estimate demand
01 · 🟢 Core
Demand forecast = estimate of future demand
🧠 What it means
A demand forecast is a time-bound estimate of future customer demand for a product or group. It is a planning input with assumptions and uncertainty, not a promise, target or historical result.
💬 OPERATOR TRANSLATION
“The number we plan around, with humility built into the spreadsheet.”
🛍️ In real life
A planner estimates 900 units of a winter coat for next month, then records the assumptions behind that number.
⚠️ Watch out
Separate unconstrained demand from sales capped by stock or capacity.
🔗 Related: Demand planning · Constrained sales forecast · Seasonality · ↑ all terms
02 · 🔵 Operations
Demand planning = coordinated decisions about future demand
🧠 What it means
Demand planning combines forecasts, commercial inputs and operational constraints to decide how an organization will prepare for expected demand. It connects a forecast to inventory, supply, finance and action.
💬 OPERATOR TRANSLATION
“The meeting where Sales, Finance and Operations negotiate with the same future.”
🛍️ In real life
A retailer reviews a seasonal forecast with marketing, buying and supply teams before placing supplier orders.
💡 Why it matters
A forecast is one estimate; a plan assigns actions, owners and assumptions.
🔗 Related: Demand forecast · Constrained sales forecast · Seasonality · ↑ all terms
03 · 🔵 Operations
Constrained sales forecast = feasible sales estimate
🧠 What it means
This forecast estimates sales the business can actually fulfill after considering limits such as inventory, production, distribution or channel capacity. It differs from unconstrained demand, which asks what customers might buy without those limits.
💬 OPERATOR TRANSLATION
“Demand wants 1,000. The warehouse has 640 and a calendar.”
🛍️ In real life
A planner forecasts interest at 1,000 units but limits the shippable sales plan to available supply.
⚠️ Watch out
Keep the unconstrained demand estimate visible so supply shortages are not mistaken for weak demand.
🔗 Related: Demand forecast · Demand planning · Seasonality · ↑ all terms
04 · 🔵 Operations
Seasonality = recurring demand pattern
🧠 What it means
Seasonality is a recurring pattern in demand associated with calendar periods or events, such as weather, holidays or recurring shopping occasions. A seasonal pattern repeats; it should be distinguished from a one-time trend or shock.
💬 OPERATOR TRANSLATION
“The umbrellas sell every rainy season. This rainstorm is not the whole forecast.”
🛍️ In real life
A garden retailer adjusts its baseline for spring demand and checks whether last year’s promotion makes the comparison misleading.
🟡 Watch out
Separate seasonality from changing trend, promotions and one-off disruptions.
🔗 Related: Demand forecast · Demand planning · Constrained sales forecast · ↑ all terms
05 · 🔵 Operations
Demand sensing = near-term forecast adjustment from recent signals
🧠 What it means
Demand sensing uses recent and relevant signals to adjust a near-term forecast. Inputs may include current orders or other timely data; the value depends on signal quality and whether the change improves decisions.
💬 OPERATOR TRANSLATION
“Yesterday’s orders can update next week. They cannot tell us next year’s weather.”
🛍️ In real life
A planner adjusts the next two weeks after order intake changes, while leaving the seasonal baseline for longer-range planning.
⚠️ Watch out
A fresh signal is not automatically informative; test it against a stable baseline and outcomes.
🔗 Related: Demand forecast · Demand planning · Constrained sales forecast · ↑ all terms
Measure forecast performance
06 · 🟢 Core
Forecast bias = persistent signed over- or underforecast
🧠 What it means
Forecast bias tracks the direction of forecast errors over time. Teams must define the sign convention; when forecast minus actual is positive, the forecast is above actual, but reporting conventions can differ.
💬 OPERATOR TRANSLATION
“The forecast is consistently optimistic, and the warehouse is consistently doing the cleanup.”
🛍️ In real life
A category repeatedly forecasts 1,200 units against 1,000 actual demand; planners inspect promotions, overrides and launch assumptions.
⚠️ Watch out
Bias shows direction; pair it with an error measure to understand the size of misses.
🔗 Related: WAPE · MAPE · FVA · ↑ all terms
07 · 🟢 Core
WAPE = total absolute error divided by total actual demand
🧠 What it means
Weighted absolute percentage error sums absolute forecast misses and divides by summed actual demand across the chosen items and periods. It weights high-volume observations more heavily and requires a clearly defined scope.
💬 OPERATOR TRANSLATION
“Big sellers count more in this report. The denominator gets a seat too.”
🛍️ In real life
Across a month, forecasts miss by 80 units in total against 1,000 actual units: WAPE is 8% for that scope.
🧮 Formula
WAPE = sum of absolute forecast minus actual differences divided by summed actual demand, using one stated unit, item set and horizon.
🔗 Related: Forecast bias · MAPE · FVA · ↑ all terms
08 · 🔵 Operations
MAPE = mean absolute percentage error
🧠 What it means
MAPE averages the absolute percentage error for each included observation. Each period contributes its percentage error equally; zero actuals make the ratio undefined, and small actuals can produce very large values.
💬 OPERATOR TRANSLATION
“A quiet Tuesday can shout as loudly as Black Friday in this average.”
🛍️ In real life
A team compares weekly percentage misses and excludes or separately handles zero-demand weeks under a documented policy.
⚠️ Watch out
State the zero-demand treatment and compare the same horizon; MAPE can exaggerate small-volume misses.
🔗 Related: Forecast bias · WAPE · FVA · ↑ all terms
09 · 🔵 Operations
FVA = change in forecast error from a planning step
🧠 What it means
Forecast value added evaluates whether a process step or input improves forecast accuracy relative to a defined baseline. A positive result requires an explicitly defined sign convention; compare the same items and horizon.
💬 OPERATOR TRANSLATION
“Before adding another committee, ask whether the last one beat the simple forecast.”
🛍️ In real life
The analyst compares error before and after a sales override for the same 200 products and planning horizon.
⚠️ Watch out
Use matched scope and state the baseline; FVA evaluates the step, not a forecaster’s worth.
🔗 Related: Forecast bias · WAPE · MAPE · ↑ all terms
Turn plans into stock decisions
10 · 🟢 Core
Replenishment lead time = wait until stock is sale-ready
🧠 What it means
Replenishment lead time is the elapsed time from initiating a replenishment until units are available for the intended use or sale. The exact start and end points should be defined; supplier production, transport and receiving can all contribute.
💬 OPERATOR TRANSLATION
“The supplier’s ship date is not the moment the product magically appears on your shelf.”
🛍️ In real life
A buyer places an order on 1 May, but counts stock as available only after production, transit and receiving finish.
⚠️ Watch out
Use the full, operationally relevant start and end events; supplier lead time alone may omit inbound handling.
🔗 Related: Cycle service level · Order fill rate · Economic order quantity (EOQ) · ↑ all terms
11 · 🔵 Operations
Cycle service level = chance of avoiding a stockout in a cycle
🧠 What it means
Cycle service level is the share or probability of replenishment cycles in which demand is met without a stockout, under the stated model and period. It concerns whether a cycle stocks out, not the share of units filled.
💬 OPERATOR TRANSLATION
“A cycle either made it through without running dry or it did not.”
🛍️ In real life
If 92 of 100 replenishment cycles meet demand without a stockout, observed cycle service level is 92% for that sample.
⚠️ Watch out
A high cycle service level does not say how many units were filled when a stockout did occur.
🔗 Related: Replenishment lead time · Order fill rate · Economic order quantity (EOQ) · ↑ all terms
12 · 🔵 Operations
Order fill rate = share of demand fulfilled from available stock
🧠 What it means
Fill rate measures how much of defined customer demand is fulfilled from stock, often by units or order lines. The exact basis and timing matter; a backordered unit may count differently by policy.
💬 OPERATOR TRANSLATION
“The customer ordered five. The box contains how many today?”
🛍️ In real life
A five-unit order ships four units immediately; the business reports its chosen unit fill rate and separately records the backorder.
⚠️ Watch out
Specify unit, line or order basis and whether later shipments count.
🔗 Related: Replenishment lead time · Cycle service level · Economic order quantity (EOQ) · ↑ all terms
13 · 🔵 Operations
Economic order quantity (EOQ) = cost-balancing order size
🧠 What it means
EOQ is a modelled replenishment quantity that balances ordering and inventory holding costs under assumptions such as stable demand and costs. Real businesses adapt it for supplier minimums, space, service targets and changing demand.
💬 OPERATOR TRANSLATION
“The mathematically tidy order meets the supplier’s very untidy minimum.”
🛍️ In real life
A buyer’s calculated EOQ is 420 units, but the supplier minimum and storage limit require a practical review.
🧮 Formula
Basic EOQ = √(2 × annual demand × order cost ÷ annual holding cost per unit); align units and assumptions.
🔗 Related: Replenishment lead time · Cycle service level · Order fill rate · ↑ all terms
Prioritize and align
14 · 🔵 Operations
ABC analysis = group stock by business value
🧠 What it means
ABC inventory analysis groups items into A, B and C classes by a chosen value measure, often annual usage value. The cutoffs and review policy are company choices, not a universal fixed percentage.
💬 OPERATOR TRANSLATION
“Give the expensive fast movers attention before alphabetizing the whole warehouse.”
🛍️ In real life
A buyer classifies products by annual consumption value, then reviews high-value A items more frequently than low-value C items.
⚠️ Watch out
State the value basis and cutoffs; ABC is a prioritization method, not a demand-variability label.
🔗 Related: XYZ analysis · S&OP · ↑ all terms
15 · 🔵 Operations
XYZ analysis = group stock by demand variability
🧠 What it means
XYZ analysis groups items by how predictable or variable demand is. Organizations choose their variability measure and thresholds; the labels complement ABC value classes rather than replace them.
💬 OPERATOR TRANSLATION
“A product that sells every week needs a different buffer from a product that sells whenever it feels like it.”
🛍️ In real life
A planner tags steady sellers X and intermittent sellers Z under the retailer’s documented variability bands.
⚠️ Watch out
XYZ thresholds differ by organization; do not assume one universal coefficient or classification.
🔗 Related: ABC analysis · S&OP · ↑ all terms
16 · 🟢 Core
S&OP = cross-functional sales and operations planning
🧠 What it means
Sales and operations planning is a recurring cross-functional process for reconciling demand, supply and financial plans over a shared horizon. It escalates trade-offs into agreed actions rather than merely publishing a forecast.
💬 OPERATOR TRANSLATION
“One plan, multiple departments, fewer surprise pallets in the parking lot.”
🛍️ In real life
A monthly S&OP review agrees to constrain a product launch after supply cannot support the original sales plan.
💡 Why it matters
Record the chosen scenario, assumptions and decisions—not just a polished consensus slide.
🔗 Related: ABC analysis · XYZ analysis · ↑ all terms
🔀 Forecast vs plan vs target
A forecast estimates a likely outcome; a plan chooses actions under constraints; a target states an intended result. One number should not silently stand in for all three.
🔀 WAPE vs MAPE
WAPE aggregates absolute errors before dividing by total actual demand, so larger-volume observations weigh more. MAPE averages observation-level percentages and can fail at zero actuals.
🔀 Cycle service level vs order fill rate
Cycle service level asks how often a replenishment cycle avoids a stockout. Fill rate asks what share of defined units, lines or orders are fulfilled.
🔀 Reorder point vs safety stock vs EOQ
A reorder point is a trigger; safety stock is a buffer; EOQ is a modelled order quantity. Inventory-policy triggers are shared with D41; this page owns forecast and cost trade-offs.
🤔 Still confused?
Follow this thread: Demand forecast → Forecast bias → Replenishment lead time → ABC analysis. That sequence moves from the basic object or relationship to the decisions and checks it supports.
Sources and scope
Primary documentation checked on 27 September 2026. Platform features and eligibility can change; examples and cartoon situations are illustrative.
